Microwave characterisation of CaF 2 at cryogenic temperatures using a dielectric resonator technique

Size: px

Start display at page:

Download "Microwave characterisation of CaF 2 at cryogenic temperatures using a dielectric resonator technique"

Adele Sanders
5 years ago
Views:

1 Journal of the European Ceramic Society 23 (2003) Microwave characterisation of CaF 2 at cryogenic temperatures using a dielectric resonator technique Mohan V. Jacob a, a *, Janina E. Mazierska, Dimitri O. Ledenyov a, Jerzy Krupka b a Electrical and Computer Engineering, School of Engineering, James Cook University, Townsville, QLD 4811, Australia b Instytut Mikroelektroniki i Optoelektroniki Politechniki Warszawskiej, Koszykowa 75, Warszawa, Poland Abstract Properties of calcium fluoride (CaF 2 ) have been well researched at UV, visible and IR range of frequencies but not at microwave frequencies. In this work we report the loss tangent and the real part of relative permittivity " r of CaF 2 measured in the temperature range K and at frequency GHz. The tan and " r were determined by measurements of the resonant frequency and the Q 0 -factor of a TE 011 mode cylindrical copper cavity with superconducting plates containing the sample under test. The measured " r of CaF 2 was found to change from to 6.505, and the tan from to when temperature was varied from 15 to 81 K. Due to the low losses CaF 2 can be useful in construction of high Q-factor microwave circuits and devices operating at cryogenic temperatures. Keywords: CaF 2 ; Microwave ceramics; Resonators; Dielectric properties PACS: Nf, nn, i, d, t, Ex, Az, Dc, j, Am 1. Introduction Calcium fluoride crystals are used in many optical applications, including mirror substrates for UV laser systems, windows, lenses and prisms for ultraviolet, visible and infrared frequencies. CaF 2 is grown by the Stockbarger technique or the Brigdman method in diameter up to about 200 mm. Calcium Fluoride (VUV grade) crystals have the transmission range from 0.19 to 7.2 mm and low refractive index from about 1.35 to 1.51 through this range. 1 IR grade Calcium Fluoride is transparent up to 12 mm. Degradation due to moisture in the atmosphere is minimal, and polished surfaces may be expected to withstand several years exposure to normal atmospheric conditions. Due to its low refractive index, Calcium Fluoride can be used without an anti-reflective coating. The maximum temperature CaF 2 can tolerate is 800 C in dry atmosphere. Low solubility and wide transmission makes this material useful for many applications, including mirror substrates for UV laser systems, windows, lenses and prisms for UV and IR applications. 1 Due to the low relative permittivity and low losses, calcium fluoride can find applications in microwave planar circuits as a substrate material. Another possibility could be a hybrid high temperature superconductor * Corresponding author. (HTS) silicon technology for microwave circuits as investigated in Refs. 2 and 3. As silicon atoms diffuse into HTS films during annealing at elevated temperatures resulting in deteriorating superconducting properties, CaF 2 was investigated for its usefulness to overcome this difficulty due to its chemical stability and structural and thermal compatibility with Si and GaAs. Contrary to extensive data available for CaF 2 at optical frequencies, there is little data on microwave properties of this material, especially at cryogenic temperatures. In this paper we present results of precise measurements of the permittivity and loss tangent of CaF 2 at cryogenic temperatures from 15 to 81 K using the dielectric resonator technique. We have used the multifrequency Transmission Mode Q-Factor (TMQF) technique 4,5 for data processing to ensure high accuracy of calculated values of " r and tan. Also the thermal expansion phenomenon of the material was taken into account in the calculations. 2. Dielectric resonator measurement method The superconducting dielectric resonator technique is a modification of the metallic dielectric resonator 6,7 and has recently been used to characterise various low loss single crystal and polycrystalline dielectric materials at microwave frequencies The Hakki-Coleman version of the dielectric resonator we used for the measurements doi: /s (03)

2 2618 M.V. Jacob et al. / Journal of the European Ceramic Society 23 (2003) superconducting and the metallic parts of the cavity respectively, A S and A M are the geometric factors of the superconducting part and metallic parts of the cavity and e is the electric energy filling factor. Geometric factors A S, A M, and e to be used in (2) were computed using incremental frequency rules as follows: 11 A S ¼!2 0 ð3þ A M ¼!2 0 ð4þ Fig. 1. Schematic of a TE 011 mode dielectric resonator. of CaF 2 is shown in Fig. 1. The resonator consisted of the copper cavity of diameter of 9.5 and 3 mm height with superconducting endplates. The CaF 2 sample was machined into a cylinder with the aspect ratio (diameter to height) equal to 1.67, with 3.1 mm height and 5.0 mm diameter. The real part of relative permittivity " r were determined from measurements of the resonant frequency of the resonator with the CaF 2 sample as the first root of the following transcendental equation 11 using software SUP12: 12 k 1 J 0 k 1 b F1 ðbþþk 2 J 1 k 1 b F0 ðbþ ¼ 0 ð1þ where: I 1 k 2 a F 0 ðþ ¼ I 0 k 2 þ K0 k 2 K 1 k 2 a I 1 k 2 a F 1 ðþ ¼ I 1 k 2 þ K1 k 2 K 1 k 2 a k 2 1 ¼!2 " r c 2 k 2 z ; k2 2 ¼ k2!2 z c 2 ; k z ¼ =L and o is the angular frequency (2pf), c is velocity of light, " o is free space permeability, " r is real relative permittivity of the sample and J 0, J 1, I 0, I 1, K 0, K 1, denote corresponding Bessel and Hankel functions. The loss tangent tan of CaF 2 was computed from the measured Q 0 -factor of the resonator on the basis of the well known loss equation, 11 namely: tan ¼ 1 1 R SS R SM ð2þ e Q 0 A S A M where Q 0 is the unloaded Q-factor of the entire resonant structure, R SS and R SM are the surface resistance of the p e ¼ " r! Computed values of the geometrical factors and the energy filling factors are given in Table 1. Values of the surface resistance HTS endplates (R SS ) and copper walls (R SM ), necessary for Eq. (2) were measured in the same copper cavity but with the sapphire rod and results are given in Section Measurements of microwave properties of CaF 2 The measurement system we used for microwave characterisation of the calcium fluoride sample is shown in Ref. 10. The system consisted of Network Analyser (HP 8722C), closed cycle refrigerator (APD DE-204), temperature controller (LTC-10), vacuum Dewar, a PC and the Hakki-Coleman dielectric resonator in transmission mode. The CaF 2 sample was grown with the Brigdman method by Ref Measurements of R SS and R SM of the Hakki-Coleman cavity As mentioned in Section 2 the surface resistances R SM and R SS of the cavity needed to be measured first. To obtain precise values of R SM we have measured S-parameters (S 21, S 11,andS 22 ) around the resonance of the Hakki-Coleman resonator with the sapphire rod and copper cavity. The measured data sets were processed Table 1 The geometrical factors and the energy filling factor of the Sapphire and CaF 2 resonators Dielectric rod CaF 2 Sapphire Frequency 29.3 GHz GHz A M 22,029 22,319 A S e ð5þ

3 M.V. Jacob et al. / Journal of the European Ceramic Society 23 (2003) with the Transmission Mode Q-Factor Technique 3,4 to obtain the loaded Q L -factor and coupling coefficients as mentioned in Section 1. The TMQF method accounts for noise, delay due to uncalibrated transmission lines and its frequency dependence, and crosstalk in measurement data and hence provides accurate values of Q L and the coupling coefficients 1 and 2. The unloaded Q 0 -factor was subsequently calculated using the exact equation, 14 Q 0 ¼ Q L ð1 þ 1 þ 2 Þ ð6þ Assuming loss tangent of the sapphire rod as 10 7, the surface resistance of copper R SM was calculated with Ref. 11 based on: 1 R SM ¼ A M e tan ð7þ Q 0 where A M is the metallic geometric factor for the copper cavity. The surface resistance R SS was measured with the sapphire rod in the copper cavity with end walls comprising of a pair of high quality YBa 2 Cu 3 O 7 thin films and calculated with Ref. 12 based on: 11 1 R SS ¼ A S R SM e tan ð8þ Q 0 A M Measured dependence of surface resistances, R SS and R SM, with temperature at frequency of 24.6 GHz are presented in Fig. 2. As measurements of calcium fluoride were performed at a frequency of approximately GHz, not at 24.6 GHz, the measured values of R SM and R SS were scaled assuming the square root frequency dependence for R SM of copper and the frequency square law for R SS of superconducting endplates Measurements of microwave properties of calcium fluoride The Hakki-Coleman resonator with HTS endplates containing the CaF 2 sample was cooled from room temperature to approximately 12 K, and the resonant frequency of GHz was obtained. The S 21, S 11 and S 22 parameters data sets around the resonance were measured as a function of increasing temperature from 13 to 81 K, and the Q 0 -factor and f res were calculated using the TMQF technique and Eq. (6) as before. The real relative permittivity " r of the calcium fluoride sample was calculated from the measured resonant frequency using Eq. (1). Variation of dimensions of the CaF 2 sample with temperature as shown in Fig. 3 were taken into consideration in the computations of " r using the temperature dependence of the linear thermal expansion coefficient after Ref. 15 as given in Fig. 4. Fig. 3. Dimensions of the CaF 2 sample at cryogenic temperatures. Fig. 2. Measured surface resistances of YBa 2 Cu 3 O 7- films and copper versus temperature. Fig. 4. Temperature coefficient of CaF 2 based on Ref. 15.

4 2620 M.V. Jacob et al. / Journal of the European Ceramic Society 23 (2003) The real part of relative permittivity " r of the CaF 2 sample measured at temperatures from 15 to 81 K is shown in Fig. 5. The " r exhibited the magnitude of approximately 6.5 and increased with the temperature by approximately 0.33%; from to The loss tangent of the CaF 2 sample was calculated using (2) from the measured unloaded Q 0 -factor. The measured temperature dependence of tan is shown in Fig. 6. The loss tangent showed an increase of 86% in the temperature range from 15 to 81 K; at temperatures of 15 and 81 K, the measured tan of CaF 2 was and , respectively. Using a linear scaling, the calculated loss tangent of CaF 2 at temperature of 15 K and frequency of 10 GHz is only Our results show that CaF 2 exhibits very low losses at microwave frequencies and cryogenic temperatures, comparable to losses in Teflon Error analysis of measured parameters " r and tan of CaF 2 The accuracy of the measurement of the real part of permittivity using the dielectric resonator depends on the precision of the measurements of the resonant frequency and uncertainty in dimensions of the dielectric sample. We measured f res with a resolution of 1 Hz using the Network Analyser HP 8722C. Hence to assess uncertainty in " r measurements the error analysis was performed assuming the uncertainty in the dimensions of the sample of 0.2 and 0.5% using the software SUP Results of the error analysis are presented in Fig. 7 and show that the relative error r " r is approximately twice the uncertainty in dimensions. The uncertainty in the loss tangent measurements is caused by the uncertainty in the measured unloaded Q 0 -factor values, R SS, R SM and geometrical factors. The Most Probable Error (MPE) in tan of CaF 2 can be expressed after Ref. 10 as: " r r tan ¼ D e 2 þ e 1 Q 0 e tan DQ0 Q 0 2 R SS DR SS þ A S e tan þ DA S 2 R SS A S ð9þ Fig. 5. Measured real part of permittivity of CaF 2 as a function of temperature at GHz. R SM DR SM þ A M e tan þ DA # M 2 1=2 R SM A M Fig. 6. Measured loss tangents of CaF 2 as a function of temperature at GHz. Fig. 7. Most probable error in " r versus temperature for 0.2 and 0.5% uncertainty in samples dimensions.

5 M.V. Jacob et al. / Journal of the European Ceramic Society 23 (2003) crosstalk and delay due to un-calibrated cable and connectors and to ensure high precision of measurements. Calcium fluoride was found to exhibit " r varying from to 6.505, and tan from to in the temperature range from 15 to 81 K as given in Table 2. On the basis of performed error analysis we assessed the uncertainty in the measurements of " r and tan to be below 0.4 and 3%, respectively in the temperature range from 13 to 81 K. Our measurements have shown that CaF 2 is a very low loss material at cryogenic temperatures at frequency of 29 GHz. Hence, apart from optical applications, calcium fluoride can be useful in cryogenic microwave circuits where very low tan and resistance to atmospheric conditions are needed and the real relative permittivity of approximately 6.5 is adequate. Fig. 8. Most probable error in tan of CaF 2 versus temperature for varying uncertainty in Q o. Table 2 Dimensions and permittivity and loss tangent of CaF 2 at 15 and 81 k To assess errors in our measurements of tan we assumed uncertainties in R SS and R SM of 2%, and 0.5% for uncertainties in A S, A M and e. Calculated errors in measured loss tangent values of CeF 2 for assumed uncertainties in the Q 0 -factor measurements of 0, 0.5, 1 and 2% are presented in Fig. 8. The calculated MPE in tan for perfect Q 0 -factor measurements is approximately 1%. We assess the uncertainty in the Q 0 -factor of our measurement system as 1%. Hence the most probable error in tan of CaF 2 (which varied from to ) is between 1.7 and 3%. 5. Conclusions Temp. (K) CaF 2 Height (mm) Diameter (mm) Permittivity Tan (10 6 ) Frequency (GHz) The real relative permittivity and loss tangent of CaF 2 have been measured at frequency of GHz at cryogenic temperatures using the Hakki-Coleman dielectric resonator with superconducting endplates. The recently developed Transmission Mode Q-Factor technique was used for data processing to remove noise, Acknowledgements This work is done under the financial support of ARC-Large grant (A ) James Cook University. The first author acknowledges the James Cook University Post Doctoral Fellowship and the MRG. References 1. Del Mar Ventures. Available: CaF2.htm. 2. Kwon, O. W., Langley, B. W., Pease, R. F. W. and Beasley, M. R., EEE Electron Devices Lett., 1987, EDL 8, Bhagwat, S. S., Bhangale, A. R., Patil, J. M., Shirodkar, V. S., Pinto, R., Apte, P. R. and Pai, S. P., Growth of CaF 2 buffer on Si using low energy cluster beam deposition technique and study of its properties. Brazilian J. Physics, 1999, 29, Leong, K. and Mazierska, J., Accurate measurements of surface resistance of HTS films using a novel transmission mode Q-factor technique. J. Superconductivity, 2001, 14, Leong, K. and Mazierska, J., Precise measurements of the Q- factor of transmission mode dielectric resonators: accounting for noise, crosstalk, coupling loss and reactance, and uncalibrated transmission lines. IEEE Trans. MTT, 2002, 50, Kobayashi, Y. and Katoh, M., Microwave measurement of dielectric properties of low-loss materials by the dielectric rod resonator method. IEEE Trans. MTT, 1985, 33, Chen, L., Ong, C. K. and Tan, B. T. G., Amendment of cavity perturbation method for permittivity measurement of extremely low-loss dielectrics. IEEE Trans. Instrumentation and Measurement, 1999, 48, Krupka, J., Geyer, R. G., Kuhn, M. and Hinken, J., Dielectric properties of single crystal Al 2 O 3, LaAlO 3, NdGaO 3, SrTiO 3, and MgO at cryogenic temperatures and Microwave frequencies. IEEE Trans. MTT, 1994, 42, Krupka, J., Derzakowski, K., Tobar, M. E., Hartnett, J. and Geyer, R. G., Complex permittivity of some ultralow loss dielectric crystals at cryogenic temperatures. Measurement Sci. Tech., 1999, 10, Jacob, M. V., Mazierska, J., Leong, K. and Krupka, J., Microwave properties of low loss polymers at cryogenic temperatures. IEEE Trans. MTT, 2002, 50,

6 2622 M.V. Jacob et al. / Journal of the European Ceramic Society 23 (2003) Krupka, J., Klinger, M., Kuhn, M., Baranyak, A., Stiller, M., Hinken, J. and Modelski, J., Surface resistance measurements of HTS films by means of sapphire dielectric resonators. IEEE Trans. Appl. Superconductivity, 1993, 3, Krupka, J. Software in Fortran to calculate surface resistance of superconductors or copper, permittivity and loss tangent of dielectric materials, resonant frequency of differing TE and TM modes, The Institute of Materials Research and Technology, Warsaw, Poland. 14. Ginzton, E. L. Microwave measurements. McGraw Hill Book Co., New York, p Touloikian Y. S., Kirby R. K., Taylor R. E. and Lee T. Y. R. Thermal expansion nonmetallic solids. Thermophysical Properties of Matter Data Series, vol. 13, IFI/Plenum, New York, 1977.

MICROWAVE PROPERTIES OF LOW LOSS POLYMERS AT CRYOGENIC TEMPERATURES

To be published in IEEE Transactions on Microwave Theory and Techniques MICROWAVE PROPERTIES OF LOW LOSS POLYMERS AT CRYOGENIC TEMPERATURES Mohan V. Jacob a, Janina Mazierska a, Kenneth Leong a and Jerzy